---
title: "Investigate a service with labels | Grafana Labs"
description: "Learn how to break a service down by its labels to isolate an anomaly."
---

> For a curated documentation index, see [llms.txt](/llms.txt). For the complete documentation index, see [llms-full.txt](/llms-full.txt).

# Investigate a service with labels

`checkoutservice` is spiking. Labels help you see which dimension is driving that.

The **Labels** tab shows the labels for this service under the selected profile type (CPU in this case).

To investigate with labels, complete the following steps:

1. The `checkoutservice` panel has some huge spikes. Click **Labels** on the panel to explore it further.
2. In the **Search labels** field, type `region` so the **region** panel is on screen. If you already see it, click **Skip**.
3. At a glance, one region is using a lot more CPU than the other two. Click **Select** on the **region** panel.
   
   That gives each region its own panel so you can compare CPU usage side by side.
4. Look at the separate region panels. `ap-south-1` and `us-east-2` have similar CPU usage, but `eu-west-1` uses quite a bit more. That points to `eu-west-1` as the place to investigate next.

You’ve isolated the problem to a single region. Next, you’ll pinpoint the exact function responsible.